Ann. Zool. Fennici 44: 405–414 ISSN 0003-455X (print), ISSN 1797-2450 (online) Helsinki 19 December 2007 © Finnish Zoological and Botanical Publishing Board 2007

Pollen harvesting and reproductive rates in specialized solitary

Markus Franzén1 & Magnus Larsson2

1) Department of Ecology, Lund University, Ecology Building Sölvegatan 37, SE-223 62 Lund, Sweden (e-mail: [email protected]) 2) Department of Plant Ecology, Evolutionary Biology Centre, Uppsala University, Villavägen 14, SE-752 36 Uppsala, Sweden

Received 23 Jan. 2007, revised version received 29 Oct. 2007, accepted 29 Oct. 2007

Franzén, M. & Larsson, M. 2007: Pollen harvesting and reproductive rates in specialized solitary bees. — Ann. Zool. Fennici 44: 405–414.

Andrena humilis is an endangered oligolectic solitary and has declined in recent decades throughout western . The aim of this study was to explore the pollen harvesting pattern and to determine the reproductive rate in specialized andrenid bees. We measured the amount of pollen required to produce one brood-cell, the pollen harvesting rate and compared our results with data for other specialized andrenid bee species. Pollen-foraging trips were registered and the activity events (entering, leaving or digging) recorded at the nests. The mean number of pollen-foraging trips per day was 5.3 and an average bee nest was active (and open) 88 min day–1. The bees were highly efficient in harvesting pollen and spent on average 10.7 min to complete one pollen-foraging trip. Most pollen-foraging trips (77%) were completed in less than 15 min. The duration of pollen-foraging trips increased over the day, presumably because pollen became more costly to harvest. Based on pollen counts (pollen loads on bees and pollen provisions) an average bee required 3.85 foraging trips to complete one brood cell and one bee managed to accomplish 1.37 brood cells in one day with suit- able weather. In the literature we found data on an additional 19 specialized andrenid bee species. humilis seems to be extremely efficient compared with most other species, with an average trip for pollen lasting almost one hour (average for andrenid bees = 46 min). An extremely low reproductive rate seems to be a common trait among specialized bees in the family with an average 0.9 offspring produced per day and less than ten offspring produced during the whole lifetime. The high degree of specialisation and the low reproductive rate among andrenid bees can explain the severe decline in many species today.

Introduction in landscapes that provide access to multiple required resources (Michener & Rettenmeyer Most have to bridge some distances in 1956, Matheson et al. 1996). The more limited space and time to provide all resources neces- a resource becomes, the higher degree of spe- sary for survival, and many species only survive cialization is required to effectively utilize that

This article has been scientifically edited by: Johan Kotze 406 Franzén & Larsson • Ann. ZOOL. Fennici Vol. 44

rates in specialized andrenid bees. We studied the endangered and specialized solitary bee Andrena humilis Imhoff (Andrenidae). Our measurements included observations of daily bee activities, the rate of pollen harvested and the pollen required to produce one offspring. We expected the dura- tion of pollen-foraging trips to be negatively related to the time of day (when pollen becomes more costly to harvest), and female bees to differ in pollen harvesting efficiency. Our results were subsequently compared with studies of other Fig. 1. The distribution of Andrena humilis in Sweden. specialized andrenid bees. (a) until 1989 and (b) 1990–2007. Material and methods resource (Gillman & Crawley 1990). It is also advantageous to utilize a resource of high quality Study area and study species and to minimise the cost of travel (Mayr 1963, Plowright & Laverty 1984). Thus, time and This study was performed in Råshult, Sten- energy budgets are important to optimize while brohult parish in southern Sweden (56°37´N, collecting food, and therefore influence popula- 14°11´E). Råshult consists of ca. four hectares of tion survival, growth and reproduction (Stephens traditionally managed, dry meadows where the & Krebs 1986). Consequently, to maximize the pollen plant Leontodon hispidus occurs (Nils- number of offspring, foraging should be most son & Nilsson 2004). We studied the endan- efficient at times when food is abundant. gered solitary bee Andrena humilis, which is a Food resources in terms of nectar or pollen black, medium-sized bee (ca. 10 mm), special- limit the population size of many (Minck- ized on foraging pollen from species in the ley et al. 1994), and specialized solitary bees group Lactuceae () (Westrich 1990). depend on abundant pollen resources (Larsson & In the study area this bee was observed to forage Franzén 2007). Most non-parasitic solitary bees pollen mainly from L. hispidus. In Sweden A. are obligate pollen consumers and often forage humilis have been reported to forage pollen from for pollen from a limited number of plant spe- Pilosella officinarum, L. hispidus and Hypoch- cies. Each single adult female bee constructs a oeris radicata (M. Larsson unpubl. data). This nest, gathers pollen from flowers and constructs bee species has probably matched its emergence separate brood cells for each egg. Pollen-forag- to the phenology of its pollen plant, as described ing females of solitary bees have evolved mor- in similar systems (Minckley et al. 1994, Minck- phological and behavioural adaptations to forage ley et al. 1999). It nests gregariously in the pollen efficiently (Wcislo et al. 1994, Herrera ground where vegetation is sparse or absent. 1996). Specialized solitary bees generally har- In Sweden, A. humilis has faced a dramatic vest pollen more efficiently on their preferred decline during the last 50 years and is red-listed pollen-plant than do generalist bees (Strickler as endangered (Gärdenfors 2005). Currently, A. 1979, 1982). Also, bees restricted to plant spe- humilis is known from less than 10 areas in cies where pollen is available only during a short southern Sweden (Fig. 1). period of time, have a particularly high pollen harvesting rate (Giovanetti & Lasso 2005). Many andrenid bees are declining and knowl- Nest observations edge about their foraging behaviour and repro- ductive rate are essential for the conservation A total of 450 nests were found in the study area of plant–pollinator systems. In this study we in 2005. The nests were found in three separate explored the foraging behaviour and reproductive aggregations on or close to sun-exposed path- Ann. Zool. Fennici Vol. 44 • Pollen harvesting and reproductive rates in specialized solitary bees 407 ways extending through the hay meadows. We quantified as described for the pollen provision observed the bees at their nests in one of these samples. aggregations. The nest entrances were closed during the night but left opened during the day between the pollen-foraging trips. For each nest Statistical analyses observed, the observations started when the bee opened the nest entrance in the morning and The statistical analyses were based on data from continued until the nest was closed for the day. two days when the nests were observed from Each nest was uniquely labelled and one person opening until closed. A Generalised Linear could observe up to 20 nests simultaneously. On Mixed Model was used to test the duration (min) 21 June 2005 a pilot study was performed and of pollen-foraging trips in relation to the time (h) activities in 10 nests were observed. On 30 June of day and the nestcode (nestcode as a random 52 nests were labelled and observed by three per- factor, h as covariate). The time of day of nest sons and on 1 July 64 nests by four persons. The arrival (h) was categorized into periods of one nests were observed the entire day, from 08:00 hour. Tukey’s post hoc test was performed to test (before activity started) until activity ceased. for differences in hours of arrival. We were not For each nest we recorded the time for depar- able to re-identify and give the nests the same tures, arrivals and when digging occurred at the code as the day before, thus it was not possible entrance. Thirty-six observations were excluded to analyse day-effects correcting for the nest- from the statistical analyses as it was uncertain code. ANCOVA was used to test the number of at what time bees left or entered the nest. To be foraging trips (trips) of each female in one day in able to correct for possible temperature-related relation to the mean duration (min) of the pollen- effects, the temperature was measured each hour foraging trips, the activity length measured from between 9:00–13:00 on both days and the mean the time when the nest was open until closed per temperature for each day calculated. day (activity time) and the study day (study day used as a fixed factor, trips and activity time as covariates). Student’s t-test was used to see if Pollen samples there was any difference in the duration (min) of pollen-foraging trips and the number of forag- To quantify pollen provisions, nests of A. humi- ing trips (trips) during one day between the two lis were excavated. Fifteen cells that contained study days. Linear regression was used to relate pollen provisions with eggs were collected and the effect of the number of pollen-foraging trips preserved in 70% ethanol. In the laboratory, the per day to the mean duration of pollen-foraging provision samples were sonicated (KS101, Kerry trips. The relationship between the number of Ultrasonics Ltd.) for 35 min and quantified by pollen-foraging trips per day and the time when counting a known volume proportion of the the nest were opened and closed was analysed sample under a binocular microscope (the pro- using Pearson correlation. The opening and clos- cedure was modified from Kearns & Williams ing time of the nest could not be included in the 1993). In order to measure the pollen amount ANCOVA because the variables were correlated gathered per pollen-foraging trip we sampled to each other and related to the activity length 15 females at the nest site at ca. 12:00 on 30 during one day (activity time). All statistical June 2005 when they returned from a pollen- analyses were performed in SPSS 14.0. Mean foraging trip. These sampled bees were stored values are given with ±1 SE. in 70% ethanol and sonicated to release pollen from their bodies. The bees were then moved to a new jar filled with 70% ethanol and sonicated Results again. This process was repeated three times and the bees were then put under a binocular Observations of Andrena humilis microscope for visual counting of any remaining pollen grains. The pollen content in the jars was In total, 529 pollen-foraging trips were registered 408 Franzén & Larsson • Ann. ZOOL. Fennici Vol. 44

Fig. 3. The duration of one pollen-foraging trip in Andrena humilis (n = 101).

Fig. 2. Activity at the nest in a colony of Andrena median 6.00, n = 269). The duration of pollen- humilis in southern Sweden 2004. Each count refers to foraging trips increased over the day (Table 1 either when a bee enters, leaves or digs at the nest. and Fig. 4). Trip duration was shorter in the second study day (t-test: t = 0.56, p = 0.01, df = 528). The difference between the two study days and 1120 activity events (entering, leaving or can be explained by the temperature that was digging) recorded at 116 nests. Bees were active lower during the first day (mean pollen-foraging on average 88 ± 4.1 min between 8:50–13:30 duration 11.27 ± 0.44 min, median 12.0 min, n with a distinct peak at 11:30 (Fig. 2). = 252, mean temperature 18.8 °C) as compared with that during the second day (mean pollen- foraging duration 9.68 ± 0.39 min, median 10.0 Pollen harvesting min, n = 266, mean temperature 19.9 °C). There was no difference in the number of pollen-forag- The bees were highly efficient in harvesting ing trips between the days (mean day one 5.43 ± pollen and spent on average 10.7 ± 0.3 min to 0.44, mean day two 5.86 ± 0.51, t-test: t = 0.645, complete a pollen-foraging trip (range 2–35, p = 0.520, df = 90). median 9.00, n = 529). Most pollen-foraging The mean number of pollen-foraging trips per trips (77%) were completed in less than 15 min day and bee was 5.3 ± 0.33 (range 1–17, median (Fig. 3). The bees spent 7.1 ± 0.28 min inside 5.0, n = 101). The bee individuals that completed the nests between the foraging trips (range 2–34, more pollen-foraging trips per day spent less

Table 1. The duration of pollen-foraging trips in relation to the hour of nest entrance and the nestcode (random factor). The number of pollen-foraging trips was related to the duration of pollen-foraging trips by one individual, the total activity-time of one bee and the study day (fixed factor).

Dependent variable Source of variation df F P

Duration of pollen-foraging trip Nestcode 1, 357 1.47 0.015 Hour 1, 357 52.7 < 0.001 Nestcode ¥ Hour 1, 357 1.46 0.020

Number of pollen-foraging trips per day Study day 1, 88 7.65 0.007 Mean duration of tripsa 1, 88 33.0 < 0.001 Activity time 1, 88 231 < 0.001 a Mean duration of pollen-foraging trips by one individual. Ann. Zool. Fennici Vol. 44 • Pollen harvesting and reproductive rates in specialized solitary bees 409

Pollen provisions and daily harvest

On average 876 159 ± 78 141 (n = 15) pollen grains were collected during a foraging trip. The complete pollen provisions per cell consisted of on average 3 378 585 ± 134 340 pollen grains (n = 12). No evidence was found for a sex biased pollen provision. Based on this, an aver- age bee requires 3.85 foraging trips to complete one brood cell and manages to accomplish 1.37 brood cells per day.

Discussion

Highly specialized bee–flower systems are sub- Fig. 4. The duration of pollen-foraging trips in Andrena ject to strong selective forces that shape the humilis in relation to the time of the day. Error bars behaviour and morphology in bees (Minckley et show 95% CI of the mean. Error bars not sharing same al. 1999, Goulson 1999). For example, the rate letters are significantly different p( < 0.05, see Table 1). of pollen-harvesting is an important adaptive constraint in systems involving a limited pollen time per foraging trip than did individuals that resource. The solitary bee Andrena humilis is completed fewer trips, and bees that were active specialized on utilizing its pollen plants and pos- for longer completed more foraging trips than sesses a number of behavioural adaptations. In bees that were active for a shorter period of time the present study an average pollen-foraging trip (Table 1 and Fig. 5). The number of pollen-for- was only ca. 10 min and contained ca. 880 000 aging trips per day was not related to the time pollen grains. In other systems involving special- when the nest was first active during the day (r ized andrenid bees, the pollen-foraging trips have = –0.16, p = 0.16; Fig. 6a) but the number of been reported to be considerably longer (on aver- pollen-foraging trips increased the later the nest age 46 min, see Table 2) and result in a smaller was closed (r = 0.52, p < 0.001, Fig. 6b). pollen load. It is notable that pollen-foraging

Fig. 5. — a: The number of pollen-foraging trips per day in relation to the total activity period during one day (from the time the nest was open until closed) (n = 92). Number of foraging trips = 1.46 + 0.06 ¥ activity time (min) (r 2 = 0.67). — b: The number of pollen-foraging trips per day in relation to the mean pollen-foraging duration. : day 1 (n = 49); solid line: number of foraging trips = 7.45 – 0.17 ¥ time (min) (r 2 = 0.08). : day 2 (n = 43); dotted line: number of foraging trips = 11.69 – 0.52 ¥ time (min) (r 2 = 0.43). 410 Franzén & Larsson • Ann. ZOOL. Fennici Vol. 44

Fig. 6. — a: The relationship between the number of pollen-foraging trips accomplished in one day and the time when the nest were opened (n = 77). — b: The number of pollen-foraging trips in one day related to the time when the nests were closed (n = 77). Number of foraging trips = –20.93 + activity end (r 2 = 0.27).

patterns differ between solitary bee species. One harvesting, many plant taxa require a specific of the most efficient bees is Ptilothrix plumata set of characters that specialized bees need to (Anthophoridae) that spend only ca. 2.4 min on possess. For example the solitary bee Megachile a pollen-foraging trip (Schlindwein & Martins pilicrus has evolved a special hindleg brush to 2000). Other bees are reported to spend 2 h or be able to harvest pollen more efficiently from more on a single pollen-foraging trip (e.g. Geb- its exclusive pollen-plants (Müller & Bansac hardt & Röhr 1987). The high pollen harvesting 2004). Similarly, a number of bee species have rate of A. humilis is likely to be the result of an adapted to harvesting high pollen quantities from adaptation to the predictable, but time restricted, plants that present pollen during short periods of pollen resource provided by L. hispidus. the day (Table 2). For A. humilis, the duration of Plants presenting pollen over a long period pollen-foraging trips increased with time of day, of time normally attract pollen foragers through- but was distinctly higher at the very end of the out the day. For example, in the case of pollen day. Early in the day the humidity and the low foraging on Dipsacacaeae herbs ( arven- temperatures influence the duration of pollen- sis — e.g. Dasypoda argentata and Andrena foraging trips. Over the day pollen availability hattorfiana; pratensis — e.g. Andrena is believed to decrease as flower visitors of all marginata), specialized bees harvest pollen for kind collect or remove pollen. Pollen availability most of the day and are characterized by long may become particularly low near the nesting foraging trips (Table 2). Larsson and Franzén site during the day and some flowers are clos- (2007) found that the foraging trips of A. hattor- ing already around 12:00. Therefore the bees are fiana lasted ca. 58 min and result in a load of ca. forced to visit more flowers to fill their scopae 29 000 pollen grains. In contrast, solitary bees and to fly longer distances from the nest. These such as A. humilis, but also A. chalybaea and A. features may explain the observed increase in haynesi (Thorp 1969, Osgood 1989) that are spe- duration of foraging trips late in the day. Osgood cialized on plants with short pollen presentation, (1989) found a similar result for Andrena rud- are favoured by harvesting pollen rapidly and beckiae which required longer foraging trips have only brief stays in the nest between trips. towards the end of the day. This indicates that A. An ability to switch from one pollen host- humilis require dense plant populations close to plant to another has been observed in several the nesting ground to prevent increased mortality species of Andrena (Cruden 1972), but A. humi- due to prolonged foraging duration. lis was found to utilize a few pollen plants in The total number of foraging trips accom- the study area. Apparently, for efficient pollen plished in one day was negatively related to the Ann. Zool. Fennici Vol. 44 • Pollen harvesting and reproductive rates in specialized solitary bees 411 stimated values E stimated

e–g Reference Giovanetti & Lasso 2005 Linsley & MacSwain 1959 Linsley & MacSwain 1955 Larsson & Franzén 2007 Gebhardt & Röhr 1987 M. Larsson unpubl. data, Miliczky 1988 Neff & Simpson 1997 Popova 1983 Schönitzer & Klinksik 1990 Hirashima 1972 Osgood 1989 M. Larsson unpubl. data Vleugel 1947 Michener & Rettenmeyer 1956 Münster-Swendsen 1968, 1970 1968, Münster-Swendsen Thorp 1969 Davis & Laberge 1975 This paper Parker & Griswold 1982 Gebhardt & Röhr 1987 g

– – – – – – – – – – – – – 5.8 ca. 6 ca 2.5 ca. 3.6 ca. 2.5 ca. 8.3 ca. 6.7 ca. 11.2 Lifetime offspring

f

– – – – – – – – – – – 0.9 1.37 ca. 2 ca. 1 ca. 0.6 ca. 0.4 ca. 0.6 ca. 0.6 ca. 0.7 per day 0.53 ± 0.03

e

– 8 – – – – – – – – – 3.85 ca. 4 ca. 6 ca. 6 Trips c ells ca. 11 ca. 12 ca. 11 per cell 14 ± 2.5 10.4 ± 0.54 4.63 ± 0.10

d

– – – – 5.5 ca. 3 ca. 7 ca. 7 ca. 5 ca. 4 ca. 5 ca. 7 ca. 3 ca. 8 ca. 4 ca. 8 ca. 7 ca. 6 ca. 4 Trips ca. 2.4 per day 5.3 ± 0.33 Values valid for an average female bee of respective species. respective of bee female average an for valid Values a–g

c

– – – – – – 17 ca. 4 (min) ca. 20 ca 2.0 ca. 23 ca. 24 ca. 24 ca. 30 ca. 18 ca. 0.5 55 ± 9.1 9.2 ± 0.7 Nest stay 7.1 ± 0.28 11.0 ± 0.7 7.54 ± 0.56 b – 46 trip (min) ca. 44 ca. 95 ca. 65 ca. 16 ca. 30 ca. 60 ca. 14 ca. 12 ca. 12 ca. 52 ca. 3.0 ca. 170 57 ± 3.0 59 ± 3.4 38 ± 2.2 74 ± 6.7 Foraging 33.5 ± 2.1 10.7 ± 0.3 33.3 ± 2.3

a

45 82 82 360 420 464 338 450 243 350 540 563 180 610 605 570 135 188 270 585 (min) per day Foraging

spp.

spp. spp. spp. spp. Main pollen-plant species Several Ranunculus californicus Oenothera dentata Salix Salix Salix Salix Brassica Spiraea latifolia mesochoreum Unknown Hypochoeris radicata Camissonia ovata Leontodon hispidus virginica Helianthus anomalus Calluna vulgaris Several family Pollen-plant Ranunculaceae Onagraceae Salicaceae Rosaceae Dipsacaceae Liliaceae Asteraceae Salicaceae Asteraceae Onagraceae Asteraceae Portulacaceae Brassicaceae Asteraceae E ricaceae Salicaceae Salicaceae Asteraceae

ssp. (2) spp. (3) Host-plant information and reproductive traits of specialized solitary bees in the Andrenidae family. Andrenidae the in bees solitary specialized of traits reproductive and information Host-plant C umulative time (min) one bee is active searching for pollen outside her nest during day. The time (min) from the when leaving nest, until returning to nest with pollen loads filled (i.e. t he required accomplish one foraging trip). The number of pollen foraging trips one bee perform in day. The number of pollen foraging trips required to produce one brood cell (pollen provision). The number of brood cells (pollen provisions) produced during the lifetime. The time (min) a bee stay in the nest between pollen foraging trips. The number of brood cells (pollen provisions) produced in one day.

Table 2. Table from calculations based on time measurements nest observations. Species Andrena agilissima Andrena Andrena a b c d e f g Andrena clarkella Andrena crataegi Andrena erythronii Andrena polita Andrena vaga Panurgus banksianus M E AN Andrena chalybaea Andrena humilis Andrena erigeniae Panurginus crawfordi Andrena haynesi Andrena fuscipes Andrena erythrogaster Andrena nycthemera Andrena rudbeckiae 412 Franzén & Larsson • Ann. ZOOL. Fennici Vol. 44 duration of the pollen-foraging trips of individu- to forage close to the nest (Naef-Daenzer 2000, als and positively related to the activity length Williams & Tepedino 2003). In fragmented land- during one day. Possibly, some individuals con- scapes an optimal forager must decide how long centrate on pollen-foraging during certain days, to stay in a patch and where to continue its search or perhaps some individuals are much more if it leaves the patch. Increased habitat destruc- effective than others. Linsley and MacSwain tion will especially affect species using multiple (1959) suggested that the number of trips per and spatiotemporally separated resources such day are related to the state of the nest and that as flower-visiting bees and (Cane 2001, bees accomplish more trips to complete unfin- Ewers & Didham 2006). A low fecundity and ished brood cells. The total number of foraging thus a low reproductive rate seem to be a gen- trips accomplished in one day was related to the eral pattern among several bees and influences activity length but not the time of nest opening. population size and extinction risks (cf. Hanski The pollen resource is probably superabundant & Singer 2001). Further studies are required to only for a short period of the day. During the compare different traits in specialized vs. gener- peak of bee activity at 11:30 the pollen-forag- alist bees. ing trips were the shortest. This suggests that From a conservation point of view it is the factor limiting A. humilis populations is important to highlight that some bees are active the time of pollen presentation rather than the during extremely short periods and are, thus, resource amount in terms of pollen. This short hard to study and detect. Immediate conservation pollen presentation scheme of the pollen plant L. measures are required to rescue A. humilis and its hispidus has potentially evolved to reduce pollen cleptoparasite Nomada integra from extinction loss to flower visitors and thereby increase the in several European countries, such as Sweden plant’s fitness. The timing and control of pollen (Gärdenfors 2005) and Ireland (Fitzpatrick et al. release to flower visitors reflect selection for suc- 2006). Andrena humilis may particularly suffer cessful pollination (Wilson & Thomson 1999, from a decline of L. hispidus (Nilsson & Nilsson Fenster et al. 2004). The pollination system of 2004). Bees are important pollinators and con- L. hispidus has not been studied, but the short tribute important ecosystem services and should pollen presentation might have evolved to avoid bee highlighted in conservation (Kearns et al. species such as A. humilis, that harvest large 1998). Many solitary bees may serve as umbrella amounts of pollen (cf. Westerkamp 1996). species encompassing many other insects and The number of pollen provisions made per probably indicate a long history and high habitat day and the reproductive rate depend on the quality of floral resources and nesting ground. time spent and the amount of pollen transported A. humilis is probably one example of such a during an average pollen-foraging trip. Andrena species. humilis made on average 1.4 pollen provisions per day. In the Andrenidae family one pollen Acknowledgement provision per day seems to be the norm (aver- age 0.9, Table 2). The amount of pollen carried We thank Sofia Larsson and Charlotte Jonsson for help in the during one pollen-foraging bout was also very field. We thank L. Anders Nilsson, Sven G. Nilsson, Claire high in A. humilis, thus the number of trips Holeton, Anna Persson, Henrik Smith and Erik Öckinger required to produce one offspring is low (3.8) as for fruitful discussions and comments on the manuscript. compared with that in other andrenid bee species Financial support was provided by Lunds Djurskyddsfond, (average 8, Table 2). Societas Phytogeographica Suecica and FORMAS. Johan Kotze improved the English.

Implications for conservation References

Solitary bees utilize multiple resources for Cane, J. H. 2001: Habitat fragmentation and native bees: a building nests, and harvesting pollen and nectar premature verdict? — Conservation Ecology 5: 149– (Eickwort & Ginsberg 1980), and it is beneficial 161. Ann. Zool. Fennici Vol. 44 • Pollen harvesting and reproductive rates in specialized solitary bees 413

Cruden, R. W. 1972: Pollination biology of Nemophila men- nesting habitats and flower relationships of some species ziesii (Hydrophyllaceae) with comments on the evolu- of Melandrena (). — Pan-Pacific Ento- tion of oligolectic bees. — Evolution 26: 373–389. mologist 31: 178–185. Davis, L. R. & Laberge, W. E. 1975: The nest biology Linsley, E. G. & MacSwain, J. W. 1959: Ethology of some of the bee Andrena (Ptilandrena) erigeniae Robertson Ranunculur insects with emphasis on competition for (Hymenoptera: Andrenidae). — Illinois Natural History pollen. — University of California Publications in Ento- Survey Biological Notes 95: 1–16. mology 16: 1–45. Eickwort, G. C. & Ginsberg, H. S. 1980: Foraging and Matheson, A., Buchmann, S. L., O’Toole, C., Westrich, P. & mating behaviour in Apoidea. — Annual Review of Williams, I. H. 1996: The conservation of bees. — Aca- Entomology 25: 421–446. demic Press, London. Ewers, R. M. & Didham, R. K. 2006: Confounding factors in Mayr, E. 1963: Animal species and evolution. — Harvard the detection of species responses to habitat fragmenta- University Press, Cambridge. tion. — Biol. Rev. Camb. Phil. Soc. 81: 117–142. Michener, C. D. & Rettenmeyer, C. W. 1956: The ethology Fenster, C. B., Armbruster, W. S., Dudash, M. R., Wilson, of Andrena erythronii with comparative data on other P. & Thomson, J. D. 2004: Pollination syndromes and species (Hymenoptera, Andrenidae). — The University floral specialization. — Annu. Rev. Ecol. Evol. Syst. 35: of Kansas Science bulletin 16: 645–684. 375–403. Miliczky, E. R. 1988: Observations on the bionomics of Fitzpatrick, U., Murray, T. E., Byrne, A., Paxton, R. J. & the bee Andrena (Tylandrena) erythrogaster Ashmead Brown, M. J. F. 2006: Regional Red Data List of Irish (Hymenoptera: Andrenidae) with notes on A. (Micran- bees. — Report to national parks and wildlife service drena) personata Robertson and A. (Holandrena) c. (Ireland) and environment and heritage service, Dublin. cressonii Robertson. — Illinois Natural History Survey Gärdenfors, U. 2005: The 2005 Redlist of Swedish species. Biological Notes 130: 1–18. — Swedish species information centre, Uppsala. Minckley, R. L., Cane, J. H., Kervin, L. & Roulston, T. H. Gebhardt, M. & Röhr, G. 1987: Zurbionomie der Sandbienen 1999: Spatial predictability and resource specialization Andrena clarkella (Kirby), A. cineraria (L.), A. fusci- of bees (Hymenoptera : Apoidea) at a superabundant, pes (Kirby) und ihrer Kuckucksbienen. — Drosera 87: widespread resource. — Biological Journal of the Lin- 89–114. nean Society 67: 119–147. Gillman, M. P. & Crawley, M. J. 1990: A comparative evalu- Minckley, R. L., Wcislo, W. T., Yanega, D. & Buchmann, S. ation of models of Cinnabar moth dynamics. — Oecolo- L. 1994: Behaviour and phenology of a specialist bee gia 82: 437–445. (Dieunomia) and sunflower Helianthus ( ) pollen avail- Giovanetti, M. & Lasso, E. 2005: Body size, loading capacity ability. — Ecology 75: 1406–1419. and rate of reproduction in the communal bee Andrena Müller, A. & Bansac, N. 2004: A specialized pollen-harvest- agilissima (Hymenoptera; Andrenidae). — Apidologie ing device in western palaearctic bees of the genus Meg- 36: 439–447. achile (Hymenoptera, Apoidea, Megachilidae). — Apid- Goulson, D. 1999: Foraging strategies of insects for gather- ologie 35: 329–337. ing nectar and pollen, and implications for plant ecology Münster-Swendsen, M. 1968: On the biology of the soli- and evolution. — Perspectives in Plant Ecology, Evolu- tary bee Panurgus banksianus Kirby (Hymenoptera, tion and Systematics 2: 185–209. Apidae), including some ecological aspects. — Arsskrift Hanski, I. & Singer, M. C. 2001: Extinction-colonization Kongelige Veterinaer Og Landbohojskole 51: 215–241. dynamics and host-plant choice in butterfly metapopula- Münster-Swendsen, M. 1970: The nesting behaviour of the tions. — American Naturalist 158: 341–353. bee Panurgus banksianus Kirby (Hymenoptera, Andreni- Herrera, C. M. 1996: Floral traits and plant adaptation to dae, ). — Entomologica Scandinavica 1: pollinators: a devil’s advocate approach. — In: 93–101. Lloyd, D. G. & Barrett, S. C. H. (eds.), Floral biology: Naef-Daenzer, B. 2000: Patch time allocation and patch studies on floral evolution in animal pollinated plants: sampling by foraging great and blue tits. — Animal 65–87. Chapman and Hall, New York. Behaviour 59: 989–999. Hirashima, Y. 1972: Systematic and biological studies of the Neff, J. L. & Simpson, B. B. 1997: Nesting and foraging family Andrenidae of Japan (Hymenoptera, Apoidea). behaviour of Andrena (Callandrena) rudbeckiae Rob- Part I. Biology. — Journal of the Faculty of Agriculture ertson (Hymenoptera: Apoidea: Andrenidae) in Texas. Kyushu University 12: 1–20. — Journal of the Kansas Entomological Society 70: Kearns, C. A., Inouye, D. W. & Waser, N. M. 1998: Endan- 100–113. gered mutualisms: the conservation of plant-pollinator Nilsson, S. G. & Nilsson, I. N. 2004: Biologisk mångfald i interactions. — Annu. Rev. Ecol. Evol. Syst. 29: 83–112. Linnés hembygd i Småland. 4. Kärlväxtfloran och dess Kearns, C. A. & Williams, D. 1993: Techniques for pollina- förändring i Stenbrohults socken. — Svensk Botanisk tion biologists. — University press of Colorado, Niwot. Tidskrift 98: 65–160. Larsson, M. & Franzén, M. 2007: Critical resource levels Osgood, E. A. 1989: Biology of Andrena crataegi Robertson of pollen for the declining bee Andrena hattorfiana (Hymenoptera: Andrenidae), a communal nesting bee. (Hymenoptera, Andrenidae). — Biological Conserva- — J. N. Y. Entomol. Soc. 97: 56–64. tion 134: 405–414. Parker, F. & Griswold, T. 1982: Biological notes on Linsley, E. G. & MacSwain, J. W. 1955: Observations on the Andrena (Callandrena) haynesi Viereck and Cockerell 414 Franzén & Larsson • Ann. ZOOL. Fennici Vol. 44

(Hyrnenoptera: Andrenidae). — Pan-Pacific Entomolo- Thorp, R. W. 1969: Systematics and ecology of bees of gist 58: 284–287. the subgenus Diandrena (Hymenoptera: Andrenidae). Plowright, R. C. & Laverty, T. M. 1984: The ecology and — University of California Publications in Entomology sociobiology of bumble bees. — Annual Review of Ento- 52: 1–146. mology 29: 175–199. Vleugel, D. A. 1947: Waarnemingen aan het gedrag van de Popova, L. M. [Popova, L. M.] 1983: [A contribution to the Grijze Graafbij (Andrena vaga Panz.) (Hym.). — Ento- knowledge of andrenid bees biology in the middle Volga mol. Ber. 278: 185–192. area]. — Vest. Zool. 1983: 46–50. [In Russian]. Wcislo, W. T., Minckley, R. L., Leschen, R. A. B. & Reyes, Schlindwein, C. & Martins, C. F. 2000: Competition between S. 1994: Rates of parasitism by natural enemies of a the oligolectic bee Ptilothrix plumata (Anthophoridae) solitary bee, Dieunomia triangulifera (Hymenoptera, and the flower closing beetle Pristimerus calcaratus Coleoptera and Diptera) in relation to phenologies. — (Curculionidae) for floral resources of Pavonia cancel- Sociobiology 23: 265–273. lata (Malvaceae). — Plant Systematics and Evolution Westerkamp, C. 1996: Pollen in bee-flower relations: some 224: 183–194. considerations on melittophily. — Bot. Acta 109: 325– Schönitzer, K. & Klinksik, C. 1990: The ethology of the soli- 332. tary bee Andrena nyctemera Imoff, 1866 (Hymenoptera; Westrich, P. 1990: Die Wildbienen Baden-Württembergs. Apoidea). — Entomofauna 11: 377–427. — Eugen Ulmer, Hohenheim. Stephens, D. W. & Krebs, J. R. 1986: Foraging theory. Williams, N. M. & Tepedino, V. J. 2003: Consistent mixing — Princeton University Press, New Jersey, USA. of near and distant resources in foraging bouts by the Strickler, K. 1979: Specialization and foraging efficiency of solitary mason bee Osmia lignaria. — Behavioral Ecol- solitary bees. — Ecology 60: 998–1009. ogy 14: 141–149. Strickler, K. 1982: Parental investment per offspring by a Wilson, P. & Thomson, J. D. 1999: Heterogeneity among specialist bee: does it change seasonally? — Evolution floral visitors leads to discordance between removal and 36: 1098–1100. deposition of pollen. — Ecology 72: 1503–1507.

This article is also available in pdf format at http://www.annzool.net/